diff options
Diffstat (limited to 'model.py')
| -rw-r--r-- | model.py | 13 |
1 files changed, 9 insertions, 4 deletions
@@ -310,7 +310,7 @@ class PrimedGenerator(Runner): super().__init__(model) self.cuda = cuda - def __call__(self, n_seqs, seq_len, primer): + def __call__(self, n_seqs, seq_len, primer, prime_param_a, prime_param_b, recursive): # generation doesn't work with CUDNN for some reason torch.backends.cudnn.enabled = False @@ -321,10 +321,14 @@ class PrimedGenerator(Runner): sequences = torch.LongTensor(n_seqs, n_samples) # 64-bit int frame_level_outputs = [None for _ in self.model.frame_level_rnns] + if recursive: + out_sequences = sequences + else: + out_sequences = torch.LongTensor(n_seqs, n_samples).fill_(utils.q_zero(self.model.q_levels)) + q_levels = self.model.q_levels q_min = 0 q_max = q_levels - primer_freq = 440 print("_______-___-_---_-____") print("_____________--_-_-_______") @@ -337,6 +341,7 @@ class PrimedGenerator(Runner): for i in xrange(n_samples): x[:, i] = random.triangular(q_min, q_max) def sin(x): + primer_freq = prime_param_a for i in xrange(n_samples): x[:, i] = (math.sin(i/44100 * primer_freq) + 1) / 2 * (q_max - q_min) + q_min @@ -387,8 +392,8 @@ class PrimedGenerator(Runner): sample_dist = self.model.sample_level_mlp( prev_samples, upper_tier_conditioning ).squeeze(1).exp_().data - sequences[:, i] = sample_dist.multinomial(1).squeeze(1) + out_sequences[:, i] = sample_dist.multinomial(1).squeeze(1) torch.backends.cudnn.enabled = True - return sequences[:, self.model.lookback :] + return out_sequences[:, self.model.lookback :] |